Taskcluster/Monitoring/Services

Service Tier Definitions

• Tier 1: Required for TaskCluster Platform function
• Tier 2: Insights into operations of TaskCluster Platform
• Tier 3: External infra causes task failures

Services in each tier

Tier 1

• AWS (us-east-1, us-west-1/2)
• Heroku
• Tutum
• Azure
• DockerHub (moving to AWS S3 for primary automation)
• Pulse / CloudAMQP
• Mozilla LDAP (not yet, soon)

Tier 2

• Papertrail
• Influx

Tier 3

• Hg.mozilla.org
• git.mozilla.org
• github.com
• VPN -> Balrog
• AWS not in Tier 1

Project summary

• Mak an API for events related to infrastructure status
• Emit pulse messages for events
• have an out of band status page
• Consider: pause, or stop accepting tasks or stop scheduling tasks on Tier 1 failures